Naive Bayes, LDA, and QDA Classifiers
Modeling p(x|y) instead of p(y|x) changes the parameter count, and the parameter count decides the winner.
-
-
Ridge, Lasso, and Elastic Net Regularization
One sentence covers all of them. The shape of the penalty decides everything else.
-
Logistic Regression and Generalized Linear Models
Not a classifier — a conditional probability fitted by maximum likelihood.
-
Linear Regression Assumptions and Diagnostics
The model people reach for first and interpret most confidently.